• Model the data with a generalized linear model (GLM)
• Identify Age- and APP-related genes with GLM coefficients
• Co-expression gene modules by the Age and APP phenotypes with a topology overlap matrix
• Downstream GO and KEGG
• Network and master regulators with topology-overlap-matrix and TF-targets
• Five-month APP stratification

## brain
## 2m_APP  2m_WT 4m_APP  4m_WT 5m_APP  5m_WT 6m_APP  6m_WT 
##      5      5      9     13     18     13      8     11

## retina
## 2m_APP  2m_WT 5m_APP  5m_WT 6m_APP  6m_WT 
##      4      5      3      5      5      6

\(y \sim age + group + batch + age * group\)

load(file = "markdown/fit_app.rdt")
fit.app$coefficients

##                   Estimate Std. Error     t value     Pr(>|t|)
## (Intercept)    10.49802829 0.01860672 564.2062503 8.932371e-97
## age4m          -0.01065783 0.02362574  -0.4511110 6.538600e-01
## age5m          -0.03710797 0.02189447  -1.6948560 9.632304e-02
## age6m          -0.06244793 0.02436192  -2.5633422 1.342184e-02
## groupAPP        0.96882037 0.02749563  35.2354293 5.690012e-37
## batchmouse     -0.04023277 0.01993577  -2.0181194 4.896016e-02
## age4m:groupAPP  0.04023281 0.03526372   1.1409121 2.593418e-01
## age5m:groupAPP  0.05077686 0.03259252   1.5579297 1.255577e-01
## age6m:groupAPP  0.02422343 0.03673571   0.6593974 5.126664e-01

##        groupAPP age4m:groupAPP age5m:groupAPP age6m:groupAPP
## Zfp930    FALSE          FALSE          FALSE          FALSE
## Zfp941     TRUE          FALSE           TRUE           TRUE
## Znhit1    FALSE          FALSE          FALSE          FALSE

##        age4m age5m age6m age4m:groupAPP age5m:groupAPP age6m:groupAPP
## Zfp930 FALSE FALSE FALSE          FALSE          FALSE          FALSE
## Zfp941 FALSE FALSE  TRUE          FALSE           TRUE           TRUE
## Znhit1 FALSE FALSE FALSE          FALSE          FALSE          FALSE

profile.Id

##  [1] "FALSE-TRUE-TRUE-FALSE"  "TRUE-FALSE-FALSE-TRUE" 
##  [3] "TRUE-TRUE-FALSE-FALSE"  "FALSE-FALSE-TRUE-TRUE" 
##  [5] "FALSE-TRUE-FALSE-TRUE"  "TRUE-FALSE-TRUE-FALSE" 
##  [7] "FALSE-FALSE-TRUE-FALSE" "TRUE-FALSE-TRUE-TRUE"  
##  [9] "TRUE-FALSE-FALSE-FALSE" "FALSE-FALSE-FALSE-TRUE"

Gene number: 61

##   KEGGID       Pvalue                      Term
## 1  05010 0.0002631084       Alzheimer's disease
## 2  03010 0.0028387608                  Ribosome
## 3  00190 0.0054019260 Oxidative phosphorylation
## 4  04744 0.0095038471         Phototransduction

Gene number: 173

##   KEGGID       Pvalue                            Term
## 1  03010 1.311303e-08                        Ribosome
## 2  04020 1.394302e-03       Calcium signaling pathway
## 3  00330 4.113916e-03 Arginine and proline metabolism
## 4  05322 9.474179e-03    Systemic lupus erythematosus
## 5  04720 9.798741e-03          Long-term potentiation

Gene number: 179

The 278 genes that separate the 2 groups

The 278 genes that separate the 2 groups

61 GLM APP genes

133 GWAS AD genes and 61 GLM APP genes

##                    Estimate Std. Error      t value     Pr(>|t|)
## (Intercept)     5.431993400 0.02902422 187.15384094 7.932439e-73
## age4m           0.050087235 0.03685327   1.35909889 1.802143e-01
## age5m          -0.017508569 0.03415270  -0.51265546 6.104498e-01
## age6m          -0.072281386 0.03800162  -1.90206074 6.293165e-02
## groupAPP        0.005948535 0.04288983   0.13869336 8.902496e-01
## batchmouse     -0.058435065 0.03109737  -1.87909967 6.606587e-02
## age4m:groupAPP -0.002872924 0.05500710  -0.05222823 9.585549e-01
## age5m:groupAPP  0.006968397 0.05084036   0.13706428 8.915304e-01
## age6m:groupAPP  0.137731629 0.05730323   2.40355769 1.998603e-02

• Objective: identify gene groups with similar expression patterns
• Modified from WGCNA
• Build topoloy-overlap-matrix (TOM) as similarity matrix from gene expression's pearson correlation
• Dynamic clustering with the TOM
• Identify modules's signature profile
• Downstream GO and KEGG on modules

## branch
##   0   1   2   3   4   5   6   7   8 
##  99 191 159 100  69  61  58  56  49

##  [1] "2900079G21Rik" "Abhd14b"       "Acat2"         "Aldh1a2"      
##  [5] "App"           "Arpp21"        "Arrdc1"        "BC051226"     
##  [9] "Cd97"          "Chordc1"       "Creld2"        "Cryab"        
## [13] "Cys1"          "Dhcr24"        "Dnajb11"       "Dnajc3"       
## [17] "E130307A14Rik" "Fdft1"         "Fdps"          "Gm6741"       
## [21] "Gm7120"        "Gstk1"         "Hsp90b1"       "Hspa1b"       
## [25] "Hspa5"         "Hspb1"         "Insig1"        "Iqgap2"       
## [29] "Klhl29"        "Lamr1-ps1"     "Lime1"         "Magt1"        
## [33] "Manf"          "Mapk12"        "Mas1"          "Mical1"       
## [37] "Mllt3"         "Nsdhl"         "Pdia3"         "Pdia4"        
## [41] "Pdia6"         "Prnp"          "Psen1"         "Pstpip2"      
## [45] "Rab34"         "Rps6ka5"       "Sars2"         "Sc4mol"       
## [49] "Sdf2l1"        "Sh3rf3"        "Simc1"         "Slc6a13"      
## [53] "Spice1"        "Srebf1"        "Stard4"        "Stx2"         
## [57] "Taf6l"         "Tnni1"         "Tox"           "Wdr6"         
## [61] "Xbp1"          "Zdhhc14"       "Zfp316"

## [1] "Mllt3"  "Srebf1" "Taf6l"  "Xbp1"

63 Module 6 genes

133 GWAS AD genes and 63 Module 6 genes

##                   Estimate Std. Error    t value     Pr(>|t|)
## (Intercept)     4.55777975 0.04624779 98.5512945 6.119811e-59
## age4m          -0.04252858 0.05872277 -0.7242265 4.723020e-01
## age5m          -0.10883739 0.05441963 -1.9999658 5.095089e-02
## age6m          -0.09890929 0.06055257 -1.6334448 1.086576e-01
## groupAPP       -0.14849526 0.06834155 -2.1728400 3.455778e-02
## batchmouse      0.08846152 0.04955121  1.7852547 8.028640e-02
## age4m:groupAPP  0.05468232 0.08764947  0.6238751 5.355457e-01
## age5m:groupAPP  0.10056666 0.08101009  1.2414090 2.202471e-01
## age6m:groupAPP  0.10582645 0.09130817  1.1590031 2.519604e-01